Extension of the MPEG-7 Fourier Feature Descriptor for face recognition using PCA

Abstract

The Principal Component Analysis or the eigenface technique provides a practical solution to the problem of face recognition. Recently, many face descriptors for MPEG-7 have been proposed for face retrieval in video streams. In this paper, a new method for face recognition is presented based on extracting the most discriminant features of the MPEG-7 Fourier Feature Descriptors of the face space, defined by MPEG-7 face recognition technique, through the implementation of the eigenface technique. It will be demonstrated that the proposed method improves the recognition rate and copes better with pose variations under different facial expressions and varying face conditions, as well as illumination variations. In addition, the proposed method achieves substantial savings in the computation time needed by the recognition system.